In general, individuals with high reputation are more likely to be noticed. Moreover, the society also has different evaluation tendencies towards the positive or negative behaviors of high-reputation individuals. Motivated by this reality, this paper develops spatial public goods game model from three perspectives, which involve a dynamic reputation threshold based on local reputation and global reputation, heterogeneous evaluation of individual reputation, reputation-based method for selecting the target neighbor for strategy learning. Numerical experiments indicate highly positive evaluation on the cooperation strategy of individuals with high reputation always favors cooperation. And highly negative evaluation of the defection strategy of individuals with high reputation can promote cooperation under strong dilemma, while leniently negative evaluation of the defection strategy of individuals with high reputation is conducive to cooperation under weak dilemma. For different tendencies in reputation evaluation, the learning mechanism that individuals preferentially select individuals with high reputation as strategy learning objects is beneficial for promoting cooperative behavior of the system. In the strong dilemma environment, the proportion of attention to local average reputation and global average reputation has different effects on the cooperative behavior of the system under different evaluation tendencies for high reputation individuals.

中文翻译：

---

基于声誉的异构评价与学习对空间公共物品博弈合作的影响

一般来说，声誉高的人更容易受到关注。而且，社会对高声誉个体的积极或消极行为也存在不同的评价倾向。受此现实启发，本文从三个角度开发了空间公共物品博弈模型，包括基于局部声誉和全局声誉的动态声誉阈值、个体声誉的异构评估、基于声誉的策略学习目标邻居选择方法。数值实验表明，对高声誉个体合作策略的高度积极评价总是倾向于合作。对高声誉个体的背叛策略高度负面评价有利于强困境下的合作，而对高声誉个体的背叛策略宽松的负面评价则有利于弱困境下的合作。针对不同的声誉评估倾向，个体优先选择声誉高的个体作为策略学习对象的学习机制有利于促进系统的合作行为。在强困境环境下，对高声誉个体不同评价倾向下，对局部平均声誉和全局平均声誉的关注比例对系统的合作行为有不同的影响。